The long term objective is to test whether the high resolution structure of a multi-subunit protein complex can be determined by building up the complex from structures of its individual subunits and subcomplexes. We plan to investigate the membrane protein LH1 in detergent micelles by liquid NMR methods. LH1 is well suited for this study because it easy forms subcomplexes and individual subunits upon increasing the detergent concentration. Also, by comparing the structures of the individual subunits when isolated with those in subcomplexes, we will be able to test the two- stage model of membrane protein folding. To achieve these goals, the structure of the beta subunit, beta--BCh1 (B777) subcomplex, and alpha- beta-2BCh1 (B820) subcomplex will be determined by multidimensional NMR in detergent micelles. In addition, we will investigate the entire LH1 complex (B877) in detergent micelles by TROSY methods. The information from these experiments combined with biochemical constraints will be used to model the LH1 structure and ultimately determine if the structure of a multi-subunit complex can be determined from the sum of its parts. Finally, mapping the structural differences between the different I3 subunit association states will give insight into the two-stage model of protein folding.